Search

Solids holdup as one of the main parameters in characterizing the performance of fluidized bed reactors is widely concerned. With its development and improvement, visualization technology has been applied in fluidization because of its little disturbance to the flow. In this study, four types of particles with different properties are tested in a narrow rectangular fluidized bed equipped with a high-speed video camera. Calibration curves of these different types of particles are achieved by correlating the grayscale of the digital images with the corresponding solids holdup. These calibration curves are further applied to obtain the average solids holdup across the sectional area and local solids holdup from the center towards the wall in both a gas-solids turbulent fluidized bed and a circulating fluidized bed to verify the results. The calibration method works well for solids holdup of different types of particles in both dense and dilute fluidization systems. This method is important to characterize the fluidization quality and reactor performance with a wide operating condition. [ABSTRACT FROM AUTHOR]

Previous simulations of different-sized MTO fluidized beds indicated that the reactive predictions gradually deviate from the experiment with the reactor scale-up. This study aimed to find the underlying reasons by investigating effects of kinetic models and coke content. Two kinetic models were separately integrated into the multi-fluid model, and a reactor model was then established considering coke distribution. It showed that the ratio of ethylene to propylene is still over-predicted even the kinetic model considers the conversion of ethylene to propylene if using average coke content in simulation, but considering a reasonable coke content distribution both models can give better predictions in methanol conversion, hydrocarbon products and selectivity to light olefins. It indicated that coke content distribution plays a critical role in reaction predictions and the kinetic models with similar component classification mainly affect the variation of products along the bed height due to different reaction routes. [Display omitted] • The role of kinetic models are investigated in reactive simulation of an industrial MTO reactor. • The coke content plays a more critical role in reaction predictions. • The kinetic model mainly affects the variation of products along space position in reactor. • Both kinetic models considering coke distribution can give reasonable reaction predictions. [ABSTRACT FROM AUTHOR]

[Display omitted] • Quantitative analysis of the cluster characteristics in gas–solid turbulent fluidized bed by image processing. • Volume/mass fraction of each particle phase are characterized based on area statistics. • Distribution of the particle cluster characteristics are obtained under wide operating conditions. • Cluster cores movement brings a good solids mixing in the lateral direction. • Most cluster cores moved downward in turbulent fluidized bed. A two-step Otsu's method was applied to process the images of gas-solids flow field at multiple axial levels in a TFB with a H 0 / D ratio from 0.5 to 2.0. The flow field was segmented to a discrete particle phase, a dense particle phase, and a cluster core and a cluster cloud phases. The solids holdup, volume and mass fractions, and the velocity vector field of each phase were analyzed by postprocessing. The frequency, size, shape, and velocity of the cluster cores were also extracted. The increasing H 0 / D ratio results in an increase in solids holdup in each phase and a decrease in the mass and volume fraction of cluster clouds. A strong disturbance in the local flow field was resulted from the formation and movement of cluster cores and a more dynamic velocity change was found inside cluster clouds. The clusters were found to travel downwardly and with equal chances to move towards the center and the wall of the TFB at a low H 0 / D ratio with a high U g. [ABSTRACT FROM AUTHOR]

[Display omitted] • A population balance model is proposed for describing the evolution of coke content. • The PBM and multiscale CFD are efficiently combined to simulate an industrial-scale MTO reactor. • The primary light olefins and their selectivity are better predicted. • The relaxation time for expressing the effects of induction period in catalysts are considered. The coke distribution over shape-selective catalysts in fluidized bed reactors is critical to the selectivity of target products. A population balance model (PBM) is developed in this study to describe the evolution of coke distribution due to both reactions and circulation of catalysts. The PBM is coupled with multiscale computational fluid dynamics (CFD) to simulate a methanol-to-olefins fluidized bed reactor. A wide, uneven distribution of coke content can thus be predicted. The mass fractions of the desired products, ethylene and propylene, show better agreement with the experimental data than those without PBM. And the selectivity ratio of ethylene to propylene is also better predicted. Furthermore, the relaxation time of fresh catalyst is considered to express the effects of the induction period in MTO reactions. The reasonable prediction suggests that coupling PBM with multiscale CFD is helpful to understand MTO reactors and may find more application with respect to its optimization and scale-up. [ABSTRACT FROM AUTHOR]